Dust Puts Enterprise AI Agents on a Multiplayer Operating System
Dust’s USD 40 million (approx. RM184 million) Series B underscores how quickly enterprise AI agents are maturing from novelty tools into core infrastructure. Framed as a “multiplayer” operating system, Dust lets organisations deploy, orchestrate and govern fleets of specialised agents that share the same context, notifications, artefacts and goals as human teams. Instead of one-off prompts in isolated chat windows, employees and agents work together in a shared workspace that supports projects, conversations, to‑dos and document generation. An intelligence layer connects more than 100 data sources and enterprise tools so agents can act on real company knowledge rather than generic internet text. Governance features such as granular permissions, usage monitoring, audit trails and analytics make this agent fabric palatable to risk-conscious CIOs. The result is an AI platform scaling strategy focused on collaboration, observability and compliance rather than stand‑alone chatbots.
Usage Metrics Signal AI Agent Deployment at Production Scale
Adoption data from Dust suggests that enterprise AI agents are no longer confined to pilots. More than 3,000 organisations already use the platform, and over 300,000 agents have been deployed across its environment. Crucially, 70% weekly active usage indicates these agents are embedded in day‑to‑day workflows, not parked in innovation sandboxes. Even more telling is Dust’s report of zero churn in 2025, suggesting customers are standardising on an agent-centric operating system rather than experimenting with disposable proofs of concept. Built‑in memory and reinforcement loops allow agents to learn preferences and continuously improve, while SOC 2 Type II certification and GDPR compliance address data‑handling concerns. Together, these features point to an inflection point in AI agent deployment: businesses are consolidating fragmented experiments into shared, governable platforms that can scale across departments and use cases.
Hark and Overwatch AI Highlight Vertical Expansion of Agent Platforms
While Dust targets broad enterprise collaboration, other well‑funded players show how AI agents are moving into specialised verticals. Hark has raised more than USD 700 million (approx. RM3.22 billion) in a Series A round at a USD 6 billion (approx. RM27.6 billion) valuation to build AI systems and AI‑native hardware that serve as a more personal interface between humans and machines. Its vision extends beyond chatbots, combining models, software and devices that retain context and adapt to individual users. At the same time, Overwatch AI has secured USD 1.5 million (approx. RM6.9 million) in pre‑seed funding to embed an AI platform directly into aviation operations. By unifying fragmented airline systems and documentation behind natural language queries, it gives pilots and crew real‑time, compliant guidance grounded in official sources. Both startups reflect how AI platform scaling is spreading into deeply specialised, high‑stakes environments.
From Single-Player Assistants to Networked Agent Ecosystems
Taken together, Dust, Hark and Overwatch AI illustrate a broader shift from single‑player assistants to networked ecosystems of enterprise AI agents. Dust’s multiplayer operating system brings shared context and governance to thousands of organisations, Hark is reimagining the human–machine interface through tightly integrated models and hardware, and Overwatch AI shows how domain‑specific agents can cut through operational complexity in aviation. These platforms share common design principles: durable memory, deep data and tool integrations, transparent sourcing, and robust controls. They also indicate that enterprise AI is moving well beyond experimentation; investors are backing infrastructure meant to support long‑lived, mission‑critical deployments, and customers are sticking with platforms that demonstrably reduce friction and compound knowledge. As these systems mature, the competitive advantage will likely shift from who has access to models to who can orchestrate secure, collaborative swarms of agents across the entire organisation.
